GNSS-acoustic (GNSS-A) positioning technique needs to conduct in-suit sound speed profile (SSP) measurements, leading to a high cost and limiting to real-time applicability of this technique. To tackle this issue, a SSP inversion model based on a single-exponential empirical temperature profile (SETP) has been developed just recently. This contribution was to propose a novel SSP inversion model based on a double-exponential temperature profile (DETP) to improve the inversion precision. In addition, the proposed inversion model was appended with a prior constraints constructed by the marine environment product. Using the Japanese long-term seafloor geodetic observations, the superior of the proposed inversion model was validated sufficiently. The SSP inversion precision was evaluated by the in-suit SSP. It showed that the root mean square error (RMSE) of the DETP-based SSP inversion result of the whole water layer was 5.54 m/s, better than 6.92 m/s of the SETP-based SSP model, particularly in shallow and middle water layers. For water columns not exceeding 300 m depth, the mean bias, standard deviation (STD), and RMSE of DETP-based SSP inversion were 1.76, 6.36, and 6.59 m/s, respectively; while those of SETP-based SSP inversion were 2.03, 7.94, and 8.19 m/s, respectively. For the water columns from 300 m to 500 m, the mean bias, STD, and RMSE of DETP-based SSP inversion were 0.07, 3.18, and 3.18 m/s, respectively; while those of SETP-based SSP inversion were -2.76, 3.75, and 4.65 m/s, respectively. Moreover, the seafloor geodetic positioning based on the proposed DETP was more accurate than that based on SETP. Specifically, the positioning mean bias and STD in the horizontal direction are better than 0.2 mm and 2 mm, respectively, while those in the vertical direction are better than 3 mm and 2 cm, respectively. These indicate that the proposed DETP-based SSP can achieve a centimeter-precision-level positioning precision and improves the SSP inversion precision, significantly.
